I have a minor nitpick: referring to Australia as "relatively non-ligitious" [sic]. I've heard that this is a fallacy, and that Australia actually has more lawsuits per capita than the USA (which is considered by most Australians as being far more litigious).
As far as the rest of your post goes, I think it is right on the mark (but IANAL). Negligence involves not taking reasonable precautions against events which could be damaging to others. Whether these events involve a third party breaking the law or not is irrelevant.
It is interesting to note that the Usability Principles in this study seem to be correctly labelled: they *are* related to interface usability. However, the assessment methodology seems to be primarily measuring *transparency*.
I agree with you on this, although the study highlighted significant areas where the UI could be improved for both advanced users and novice users. Gnome isn't the only system which has usabilty flaws either -- Windows has quite a few itself. (Aside: One of my pet hates is wading through 3 levels of menus to get to programs, or minimizing all of the windows to access the desktop.) I don't mean to be complaining about Gnome -- I've used it quite a bit myself, but improvements are always a good thing. A few examples of what I mean:
Changing the font. This requires 3 different (independent) dialog boxes to completely change the font everywhere. Further, in the Sawfish WM customization section, the current font is shown something like "-helectiva-medium-r-normal-*-*-120-*-*-p-*-iso885 9-9". This is completely unnecessary and detracts from the usability for both a beginner and an advanced user. There is no reason to let the user edit this string directly - a font selection dialog box is all that is required. A sample of the current font would be useful here, though.
A number of people commented on the inconsistency of menu titles. Some menus have a "title item" up the top, and other's don't. Clicking on the title item makes the menu disappear. This doesn't really serve any usability purpose at all.
Some of the icons aren't immediately obvious. Actually, I'm not entirely convinced that icons make programs easier to use. Basically, they provide eye candy and a nice-sized target to click on. Most of the time, icons need a label or tooltip to explain what they do. (There was some usability study which I read which also said as much, but I can't find the link.)
Quite a few other people have suggested a "novice" mode (which could be integrated with Nautilis' user mode), and I agree with this. It would solve some of the problems that novice users had, such as not understanding what "man" meant.
Personally, I like many aspects of the UI in both Gnome and KDE. (They don't require you to download something like TweakUI to turn off all the annoying "features" like Windows does, either.) IMHO, the main problems with them comes about because of the modular nature of *nix systems: they can't always provide a seamless GUI to customize all the other programs on the computer. (In other words, they can't seamlessly integrate hardware config utilities and such.) If KDE, Gnome etc. can agree on a standard for such plug-ins, they would improve their usability immensely. It might also encourage a few hardware manufacturers to write some GUI configuration utilities, which would be good.
Is it possible that they didn't store the key in plaintext? (Given that they can decrypt the installation ID, I'd think that they might be able to encrypt their own key in some form or another.) Has anyone run a debugger on this thing and watched the value being pushed on the stack, or recompiled it with the supposed key and verified that it works correctly?
There's an even deeper conclusion to be reached here - instead of putting together a costly porting effort, they could do something that makes everyone happy - publish the programmer's manual to the device. DOS users get their software - subject to approval by their OS vendor - and everyone else gets to write their own, including the necessary low-level code.
This is a good point, although I believe this will become more and more rare. Companies are putting more functionality into software rather than hardware, and custom-built drivers are where these improvements are incorporated. For instance, some inkjet printer manufacturers put image manipulation code into their drivers to compensate for less than perfect color reproduction on their printers. Softmodems or Winmodems take this idea further. Would the hardware manufacturer supply information on how to do this properly as well? Not likely. In short, the product can't be used to its full potential (or not at all) unless someone writes not just a hardware driver, but a whole lot of signal processing code as well. This makes the drivers challenging to write, to say the least.
My point is that I don't think hardware manufacturers want to publish the raw specs for their products, especially if they have bugs in the hardware or require signal processing in the driver. And hardware manufacturers sometimes don't want (often inferior) 3rd party device drivers degrading the image of their product.
I don't think this is a fair discussion. Craig has to defend his company's position, and that's to make money, while all the other panel members have something to benefit from the growth of Linux and other free OSes.
I personally can't see why on earth Craig Mundie would agree to join this panel, unless he has some personal reason to. However, I also can't see what actual purpose Craig Mundie's recent speech and Microsoft's recently inflamatory (IMHO) license agreements serve. To me, they are both not just FUD, but blatant flamebait. The only people who (as far as I am aware) know of either know about alternatives to Microsoft software anyway.
The reason why no one objects to Oracle is because they're not attacking open source/free software for no real reason. Their software works with a lot of other software, both open source and not. "Other" software companies seem to have more a live an let live opinion of open source. (Something to the effect of: we're selling/licensing our software in the way that we think gives the best profit, and if other people want to do it differently, then that's their choice.)
While that may be true the MFC isn't an APPLICATION, the FSF equivalent to such a thing would be released under the LGPL and could be used in a commercial product with no problems at all other than the requirement of dynamic linking.
Actually, IIRC, the MFC licensing terms say that you may redistribute modified versions of MFC libraries. This is one area where MFC is more free than the LGPL. They give you all the source for all the libraries. I think MS recommends that you don't, and if you do, then you should not chuck it into the system directory. However, they don't stop you (in fact, they actually give instructions for rebuilding it). Can anyone confirm/deny this?
Of course, you still can't give out MFC source to anyone you want, like the LGPL allows.
When I started learning programming, Pascal was used, which was fine in the days of purely procedural programming.
. . .
I remember when I was learning Pascal, it was as if the class hit a brick wall when pointers were introduced. Imagine if learning pointers and intimate machine-dependent ("words"??) memory management was the prerequisite to larger programming concepts such as conditional statements, iteration, recursion, etc. The whole learning process would be stymied.
I believe you can get quite a way into programming before you hit pointers. Conditional statements, iteration, and recursion are all certainly possible to do without pointers. In my experience, learning Java simply "hides" the idea of pointers a bit -- sooner or later, people will realise that they didn't understand the importance in distinguishing an actual object from the references to that object. As soon as people start doing things like linked lists, they don't understand what's going on. In my opinion, it's better to do this stuff in a language without the object-orientated complexity of Java.
I did a first year CS class on introductory Java, but I had programmed in a variety of languages before (both OO and non-OO). They started off with some basic concepts of objects, such as inheritance, the difference between the class and the instances etc., but then just dropped the idea as soon as it came to writing code. This is because they wanted to teach stuff like how to do graphics (using their classes), but didn't want to teach people how objects and references worked. They gave people a class which was 100% static methods and variables which wrapped the Swing interface. So basically, they just taught procedural programming for the first few weeks until they got around to how objects are created. Further, most people didn't really understand objects until the next semester's course, when useful data structures were introduced and the reasons for OOP became more apparent.
The last point is that it made it more difficult (IMHO) to teach C and assembler, because there were so many new concepts that students can't really link in with their existing programming knowledge. For people who have learnt programming with the mindest that using objects is the One and Only True Way to program, doing without them completely can be a bit of a loss.
In summary: I think it's best to teach a language like Pascal (or even elementary C) so people get the idea of programming, then get people accustomed to pointers and data structures, and then show them OO languages. This makes it much easier to appreciate the reasons why OO is used in the real world.
Per that article, the Chinese government has two overriding needs: to keep their tight control over China and to embrace the Internet for economic gain. IMHO, these goals are mutually exclusive.
I was in China two years ago, and I can certainly see the Chinese government taking this stance. They have done a similar thing in regards to capitalism (which is plainly against communist ideals). Allowing people to run businesses which are independent from the government reduces their power and revenue, but is necessary for China to have a competitive economy. The Chinese government allow it (in many cases openly), but every few years they have a "crackdown" where they try to regain lost control.
My point is, it is entirely possible that China will do the same sort of thing with Internet usage. They will "regulate" it in a haphazard manner, and then when they see it posing a threat to their power, they stage a crackdown on people who run Internet cafes and don't filter sites properly. In any case, it may hasten the downfall of the government, but situations like this have existed in China for some time now (i.e. allowing private businesses and allowing tourists in).
In any case, most people in China don't seem to know or care about opinions outside of China. The government has indoctrinated them to believe that all the evils in this world originate from western society, and that China is working hard to do the best for its people. Example: the Chinese tour guides thought that westerners all knew about Tianamen Square because it was the largest square in Beijing.
This is exactly right. When I told one of my friends about this patch and asked her if she was going to upgrade her copy of IE, her response was basically, "no, I won't be going to any websites with malicious HTML code, and I don't use Outlook Express for email, so I don't need to". She's not computer illiterate either.
Anyway, I'll bet there's heaps of people who have heard about this patch but won't bother with it because of that sort of reasoning.
I have used Linux on and off for a few years, and I've learnt a bit about most of the important config files etc., but I must agree that very basic tools like something to easily change screen resolutions is desperately needed.
I recently installed Redhat 7 on my girlfriend's machine (Redhat 6.1 wouldn't recognise her video card, modem, etc. and was full of bugs), and I the screen resolution was an absolute pain in the ass to change. I can vouch for the fact that trying to use Xconfigurator (or editing XF86Config*) is difficult at best when xdm keeps trying to start and failing. It flashes the screen about 3 times every 5 seconds and then prints an error message in the middle of whatever you're typing. Not just that, but the video card had all sorts of problems trying to use 24 or 32 bit colour. I spent about 2 hours stuffing around with it until I got it working.
The modem, sound card and one of the network cards still isn't going. Yes, I know they're dodgy integrated-with-the-motherboard components but when I tried to tell her that, her attitude was, "well, if it doesn't work, then I'm not going to buy a new computer/new components to make it work". Not unreasonable. Anyway, there are a limited number of Linux "drivers" included on the motherboard CD, but their installation is nonstandard, breaks if you're using the wrong kernel version etc. and can cause incompatibility problems. Open sourcing drivers is a nice idea, but you've got to be able to supply proprietary drivers in a standard, clean way as well.
In any case, she's decided that just getting Emacs and LaTeX for Windows is good enough.
My point is: It's the hardware configuration that annoys me most about the current state of Linux development. I can't really say that I'm all that impressed with the configuration tools in Redhat at least. The tools are annoying, buggy, don't do what I want and are hard to find a lot of the time, due to a lack of centralised documentation. IMHO, Xconfigurator is a case in point. Further, the tools break as soon as you try to edit a config file manually to get the functionality you want.
There are a lot of things about Linux that I appreciate and enjoy very much. It has a wide variety of programs for all tastes, and is generally cooler in many respects than Windoze. However, I also would like a certain standard of ease of use in a few critical areas.
Before I get a whole lot of flames in reply, I'd just like to say that (a) I am a programmer (and hate VB and the like), and (b) I have spent a reasonable amount of time reading docs (far far more than the average user would bother with).
I agree with the Xenex (and Microsoft's opinion of the state of broadband in Australia). Pretty much the only option for many of us Brisbanites as far as broadband goes is cable, which is still too expensive for the average user and doesn't reach a lot of homes or businesses. ADSL is about as accessible to most people here as.NET is.
I can see exactly why Microsoft want broadband access - they've been using developers as guinea pigs for some time now. Most of the major downloads from MSDN require "Windows Installer", and over a 33.6K modem it's terrible. I downloaded 10 MB for the Platform SDK (about 3/4 hr) and then it crashes when it's downloaded without installing anything. I think the idea is to stop you from getting complete downloaded files or executables on your hard drive, so you can't redistribute it. Anyway, I really can't see this sort of thing being the ideal situation that Microsoft envisages for.NET or Whistler or whatever. No user in their right mind would put up with this.
It also means that users can probably send quite a bit of data back too, if that would ever be used...
Slashdot Mirror
As far as the rest of your post goes, I think it is right on the mark (but IANAL). Negligence involves not taking reasonable precautions against events which could be damaging to others. Whether these events involve a third party breaking the law or not is irrelevant.
I agree with you on this, although the study highlighted significant areas where the UI could be improved for both advanced users and novice users. Gnome isn't the only system which has usabilty flaws either -- Windows has quite a few itself. (Aside: One of my pet hates is wading through 3 levels of menus to get to programs, or minimizing all of the windows to access the desktop.) I don't mean to be complaining about Gnome -- I've used it quite a bit myself, but improvements are always a good thing. A few examples of what I mean:
Quite a few other people have suggested a "novice" mode (which could be integrated with Nautilis' user mode), and I agree with this. It would solve some of the problems that novice users had, such as not understanding what "man" meant.
Personally, I like many aspects of the UI in both Gnome and KDE. (They don't require you to download something like TweakUI to turn off all the annoying "features" like Windows does, either.) IMHO, the main problems with them comes about because of the modular nature of *nix systems: they can't always provide a seamless GUI to customize all the other programs on the computer. (In other words, they can't seamlessly integrate hardware config utilities and such.) If KDE, Gnome etc. can agree on a standard for such plug-ins, they would improve their usability immensely. It might also encourage a few hardware manufacturers to write some GUI configuration utilities, which would be good.
Is it possible that they didn't store the key in plaintext? (Given that they can decrypt the installation ID, I'd think that they might be able to encrypt their own key in some form or another.) Has anyone run a debugger on this thing and watched the value being pushed on the stack, or recompiled it with the supposed key and verified that it works correctly?
This is a good point, although I believe this will become more and more rare. Companies are putting more functionality into software rather than hardware, and custom-built drivers are where these improvements are incorporated. For instance, some inkjet printer manufacturers put image manipulation code into their drivers to compensate for less than perfect color reproduction on their printers. Softmodems or Winmodems take this idea further. Would the hardware manufacturer supply information on how to do this properly as well? Not likely. In short, the product can't be used to its full potential (or not at all) unless someone writes not just a hardware driver, but a whole lot of signal processing code as well. This makes the drivers challenging to write, to say the least.
My point is that I don't think hardware manufacturers want to publish the raw specs for their products, especially if they have bugs in the hardware or require signal processing in the driver. And hardware manufacturers sometimes don't want (often inferior) 3rd party device drivers degrading the image of their product.
I personally can't see why on earth Craig Mundie would agree to join this panel, unless he has some personal reason to. However, I also can't see what actual purpose Craig Mundie's recent speech and Microsoft's recently inflamatory (IMHO) license agreements serve. To me, they are both not just FUD, but blatant flamebait. The only people who (as far as I am aware) know of either know about alternatives to Microsoft software anyway.
The reason why no one objects to Oracle is because they're not attacking open source/free software for no real reason. Their software works with a lot of other software, both open source and not. "Other" software companies seem to have more a live an let live opinion of open source. (Something to the effect of: we're selling/licensing our software in the way that we think gives the best profit, and if other people want to do it differently, then that's their choice.)
Actually, IIRC, the MFC licensing terms say that you may redistribute modified versions of MFC libraries. This is one area where MFC is more free than the LGPL. They give you all the source for all the libraries. I think MS recommends that you don't, and if you do, then you should not chuck it into the system directory. However, they don't stop you (in fact, they actually give instructions for rebuilding it). Can anyone confirm/deny this?
Of course, you still can't give out MFC source to anyone you want, like the LGPL allows.
I believe you can get quite a way into programming before you hit pointers. Conditional statements, iteration, and recursion are all certainly possible to do without pointers. In my experience, learning Java simply "hides" the idea of pointers a bit -- sooner or later, people will realise that they didn't understand the importance in distinguishing an actual object from the references to that object. As soon as people start doing things like linked lists, they don't understand what's going on. In my opinion, it's better to do this stuff in a language without the object-orientated complexity of Java.
I did a first year CS class on introductory Java, but I had programmed in a variety of languages before (both OO and non-OO). They started off with some basic concepts of objects, such as inheritance, the difference between the class and the instances etc., but then just dropped the idea as soon as it came to writing code. This is because they wanted to teach stuff like how to do graphics (using their classes), but didn't want to teach people how objects and references worked. They gave people a class which was 100% static methods and variables which wrapped the Swing interface. So basically, they just taught procedural programming for the first few weeks until they got around to how objects are created. Further, most people didn't really understand objects until the next semester's course, when useful data structures were introduced and the reasons for OOP became more apparent.
The last point is that it made it more difficult (IMHO) to teach C and assembler, because there were so many new concepts that students can't really link in with their existing programming knowledge. For people who have learnt programming with the mindest that using objects is the One and Only True Way to program, doing without them completely can be a bit of a loss.
In summary: I think it's best to teach a language like Pascal (or even elementary C) so people get the idea of programming, then get people accustomed to pointers and data structures, and then show them OO languages. This makes it much easier to appreciate the reasons why OO is used in the real world.
Per that article, the Chinese government has two overriding needs: to keep their tight control over China and to embrace the Internet for economic gain. IMHO, these goals are mutually exclusive.
I was in China two years ago, and I can certainly see the Chinese government taking this stance. They have done a similar thing in regards to capitalism (which is plainly against communist ideals). Allowing people to run businesses which are independent from the government reduces their power and revenue, but is necessary for China to have a competitive economy. The Chinese government allow it (in many cases openly), but every few years they have a "crackdown" where they try to regain lost control.
My point is, it is entirely possible that China will do the same sort of thing with Internet usage. They will "regulate" it in a haphazard manner, and then when they see it posing a threat to their power, they stage a crackdown on people who run Internet cafes and don't filter sites properly. In any case, it may hasten the downfall of the government, but situations like this have existed in China for some time now (i.e. allowing private businesses and allowing tourists in).
In any case, most people in China don't seem to know or care about opinions outside of China. The government has indoctrinated them to believe that all the evils in this world originate from western society, and that China is working hard to do the best for its people. Example: the Chinese tour guides thought that westerners all knew about Tianamen Square because it was the largest square in Beijing.
Anyway, I'll bet there's heaps of people who have heard about this patch but won't bother with it because of that sort of reasoning.
I recently installed Redhat 7 on my girlfriend's machine (Redhat 6.1 wouldn't recognise her video card, modem, etc. and was full of bugs), and I the screen resolution was an absolute pain in the ass to change. I can vouch for the fact that trying to use Xconfigurator (or editing XF86Config*) is difficult at best when xdm keeps trying to start and failing. It flashes the screen about 3 times every 5 seconds and then prints an error message in the middle of whatever you're typing. Not just that, but the video card had all sorts of problems trying to use 24 or 32 bit colour. I spent about 2 hours stuffing around with it until I got it working.
The modem, sound card and one of the network cards still isn't going. Yes, I know they're dodgy integrated-with-the-motherboard components but when I tried to tell her that, her attitude was, "well, if it doesn't work, then I'm not going to buy a new computer/new components to make it work". Not unreasonable. Anyway, there are a limited number of Linux "drivers" included on the motherboard CD, but their installation is nonstandard, breaks if you're using the wrong kernel version etc. and can cause incompatibility problems. Open sourcing drivers is a nice idea, but you've got to be able to supply proprietary drivers in a standard, clean way as well.
In any case, she's decided that just getting Emacs and LaTeX for Windows is good enough.
My point is: It's the hardware configuration that annoys me most about the current state of Linux development. I can't really say that I'm all that impressed with the configuration tools in Redhat at least. The tools are annoying, buggy, don't do what I want and are hard to find a lot of the time, due to a lack of centralised documentation. IMHO, Xconfigurator is a case in point. Further, the tools break as soon as you try to edit a config file manually to get the functionality you want.
There are a lot of things about Linux that I appreciate and enjoy very much. It has a wide variety of programs for all tastes, and is generally cooler in many respects than Windoze. However, I also would like a certain standard of ease of use in a few critical areas.
Before I get a whole lot of flames in reply, I'd just like to say that (a) I am a programmer (and hate VB and the like), and (b) I have spent a reasonable amount of time reading docs (far far more than the average user would bother with).
Firstly, for the record, I am Australian.
.NET is.
.NET or Whistler or whatever. No user in their right mind would put up with this.
I agree with the Xenex (and Microsoft's opinion of the state of broadband in Australia). Pretty much the only option for many of us Brisbanites as far as broadband goes is cable, which is still too expensive for the average user and doesn't reach a lot of homes or businesses. ADSL is about as accessible to most people here as
I can see exactly why Microsoft want broadband access - they've been using developers as guinea pigs for some time now. Most of the major downloads from MSDN require "Windows Installer", and over a 33.6K modem it's terrible. I downloaded 10 MB for the Platform SDK (about 3/4 hr) and then it crashes when it's downloaded without installing anything. I think the idea is to stop you from getting complete downloaded files or executables on your hard drive, so you can't redistribute it. Anyway, I really can't see this sort of thing being the ideal situation that Microsoft envisages for
It also means that users can probably send quite a bit of data back too, if that would ever be used...
Doug